This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Fix bitfield-- != 0 transformation in fold-const.c

Hi Roger,

> My apologies, I pulled the test case straight from the bugzilla PR
> without consideration for "small" targets.  I think its safe to
> either just completely remove the ubyte field from the sdata struct,
> i.e. for all platforms, or to pad it to a short using "ubyte:8;".
> The PR concerns overflow in a single bit field and isn't affected
> by this later structure padding.


> Consider such a patch pre-approved for mainline.  At the risk of
> over stepping my authority, I also think its extremely safe to also
> fix this testcase on the 3.4 and 3.3 branches.  Mark?  Gaby?

To backport my patch to 3.4 and 3.3, I opened a new PR 14544.

Kazu Hirata

2004-03-12  Kazu Hirata  <>

	PR other/14544
	* gcc.c-torture/execute/20040307-1.c: Remove the 24-bit

Index: 20040307-1.c
RCS file: /cvs/gcc/gcc/gcc/testsuite/gcc.c-torture/execute/20040307-1.c,v
retrieving revision 1.1
diff -u -r1.1 20040307-1.c
--- 20040307-1.c	7 Mar 2004 14:45:42 -0000	1.1
+++ 20040307-1.c	12 Mar 2004 04:22:59 -0000
@@ -11,7 +11,6 @@
     unsigned int bit5:1;
     unsigned int bit6:1;
     unsigned int bit7:1;
-    unsigned int ubyte:24;
   } sdata = {0x01};
   while ( sdata.bit0-- > 0 ) {

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]